Electronic Archive Classification Method Based on Convolutional Neural Network with Fast Text Embeddings

Nowadays, the quantity of information and data are increasing significantly thorough organizations and businesses. An electronic archive is a digital repository that stores and manages to preserves electronic documents, records, and other digital content. However, failed in extracting the large data from visual images, and also unfamiliar in recognizing the valuable information. To overcome these drawbacks proposed an ensemble layer classification method Convolutional Neural Network with Fast Text embeddings (CNN-FT). Initially the data is collected from University of California Irvine Knowledge Discovery in Databases (UCI KDD) dataset then, preprocessing is performed using the data encoding, data imputation and data balancing. The related Time-series features are extracted from the preprocessing using Principal Component Analysis (PCA), Auto-Encoder (AE) and Linear Discriminant Analysis (LDA). Finally, classification is performed by CNN with FT embeddings, while CNN classifies the local features that capture semantic relationships and nuances in documentation. The results prove that proposed CNN-FT model show better performance in terms of accuracy (98.5<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">%</sup>) when compared to previous existing method of Word to Vector (Word2Vec).